General DiscussionMy attempt at evaluating the efficiency of all the items

I WAS WAITING FOR IT
LOT OF THANKS

this makes zero fucking sense. How can a single tango be 40 times better than mek?

all you're doing is saying big items are worse than small items, slightly adjusted for winrates. pretty flawed premise.

like none of the debates you even stated exist, vlads is a great pushing item, pms is used for tanking creep aggro under tower while support pulls/ranged hero harass in lane while vanguard is gotten for the purpose of early game survivability/aggression, enabliing tower diving and preventing ganks from being considered.

and none of these bogus stats take into account hero lineup, strategy, previous items, hero gotten on, etc so they're basically made even more bogus by that. like if im getting a first item heart, thats not efficient, but if im getting it as a 3rd or 4th item, thats efficient.

you can really only compare items within items that are similar to them; drums could be compared to a casual bracer or two, or something like that, but otherwise im not really seeing it.

how can a shared tango be better than a fucking cheese dude

because it is theoretical based on his algorithm

Thanks for the feedback guys.

A shared tango is so high because it gives a lot for its mere cost of 31.25 (125/4)

Let me just remind everyone that this is for item efficiency. Think when people say branches are cost effective, and this list tries to do that. But it does assume infinite item slots. This can still be useful, say in a push strat, where on the whole of a game you get less gold than in a farm off; just pick appropriate items over ~1.3

Bot Passive, maybe it isn't useful situationally.
However, it can be used generally to know what items are good on average for most heroes.

Like medallion of courage (~2) vs crystalise (~1), maybe try medallion on that sniper?

This also can remind us how bad going for a dagon 5 is, if nothing else.

nice.

I think you don't understand how items winrate works on Dotabuff.

got it buy tangos not Mek. Better win rate

Brb stacking 99 Ironwood Branches.

Oh wait...

Edit: Hex confirmed one of the worst items in the game.

Mark dude, you should not comparer items like mek and tango since they have completely different price... But think about, isent it way more to have those tangos in the early game then a mek in the midgame?
Not getting any regen with your start items is in many cases a big fail and the lane will be doomed (more or less, defending on hove bad/good opponents are).

Makes no sense, you can't just multiply an item's winrate and its gold and expect to get a value of how efficient that item is based on winrate. It's like saying distance x area = time

Edit: Actually, i didn't read the description. You're correct, but i just don't see how useful this is. Item price and item usefulness are unrelated, so why would you get something based on gold?

Thanks for comments, and I feel this is useful. Think of this situation.

There team is a five man deathball, and I've probably only got 10 minutes until they start rolling together.
If you wanted to group and fight against their deathball, what items do I want to have the highest chance of winning?
Say I'm PA and I guesstimate I'll have about 3000 gold by then:
3000/6 slots = 500 gold per item

I go to the top of the list, look for items that cost close to that amount.
Wand, Brown Boots, ROB, Wraithband, PMS, Blades of Attack (or possibly a tp)

dagon 1 > dagon 2 > dagon 5 > dagon 3 > dagon 4

i called my math teacher, he said you don't make sense and shot himself in the head

Same as above (PA) but I want to fight at 20 mins, 6000/6 slots so say 1000 gold per item now.
Urn Of Shadows, Tranquil Boots, Medallion Of Courage, QuarterStaff, RingOfAqualia, Vit Booster

Dagon 5 in the middle maybe skewed by insane stomps.
But Dagon 1 is the most efficient which is why people say don't upgrade it.

dagon 1 most efficient - well i guess these dont matter:
1. range increase
2. mana cost decrease
3. faster cooldown
4. heroes have a limited slot of 6 item slots. 1 usually holding boots, 1 for tp, effectively 4 slots (unless bots etc)

Radiance > BKB
Necro 1 > Necro 3 > Necro 2

Your formula, largely ignores the active effects of many items.
BKB, Scythe, Dagon, Necro, Linkens, Manta rank at the bottom of the list
If a newbie follows your formula, he gonna be "efficient" for the short term but dead on the long one

Dont forget to tell your carry, "Hey 2 ultimate orbs are more efficient than a bkb. Stop buying bkb pls"

Its clear more work has to be done on the formula b4 posting it for the world to see

If I had 1000 people and told them to go for 2 ultimate orbs instead of a bkb, they would probably increase in winrate on average.
This is because people can't use bkb properly on average.

I had a look at you're dotabuff for bkb and you do use it better than average, so this is not necessarily true for you.
But for most this list is solid.

If you had an army of average skilled players that never got better, this list would show them the best items to win for a given amount of gold.

So if you are exceptional at using actives, then those items are actually higher for you. This might explain why necronomicons rank bad too, probably because most don't know you can use them in teamfights and most suck at microing them.

Isnt rapier supposed to be the most cost-efficient item?

Dude, your alghoritm is worse than my math exam, it's based on a simple idea, no way this can give a true image of the cost efficiency of items.

Winrate is exponential. Cost is not.

Yiran what do you mean by that?

And rapier is most cost effective for damage.

"If you had an army of average skilled players that never got better, this list would show them the best items to win for a given amount of gold."

There you have it, the flaw in your theory

No one even the worst of noobs, want to be called average and everyone eventually get better

By buying bkb 100 times and miss using it 100 times, eventually they will learn how to use it properly
If they follow your theory of efficiency and stick to 2 ultimate orbs only, they will not only stagnate but they will get worse if they do climb up the mmr ladder

That being said, everyone here is potential competition, GO BUY 2 ULTIMATE ORBS NOW WAHAHAHHA
Climb up the ladder and つ ◕_◕ ༽つ GIFF ME MMR つ ◕_◕ ༽つ

Regardless of people not wanting to be average, by definition, a great portion of players are average. Its just how the distribution is.

I wasn't trying to offend the average playerbase. I just made a mistake by saying 'never get better'; I meant never want to get better / not (play/practice) long enough to get better.

You need something to account for item slot limitation. Maybe you could make it 'best value items at x minutes'? But then you have to account for snowballing, being behind, what role, etc.

Lolwut you're right, that would make it more useful. Right now all it can really be used for is to get the gist of how good all the items are.
Man if I had more html/programming experience I could make a website or something.
Some checkboxes, few sliders, could be sweet.

So with this logic you should fill your slots with the items closest in price to your net worth / 6? That's the most retarded thing I have ever heard. Even Blunt makes more sense.

you just wasted all your time making these stats

Dude, this all is wrong on SO many levels, you can't even imagine.

This doesn't make much sense at all.

I can see where shawn is going, but the theory is still in its infancy

Good luck bro

It will never make sense if you tie it to win rates in dota. Some items are very good but have low win rates because they complete other items and if you don't actually complete the item it's usually because you were losing already. For example, sniper maelstrom is 46% win rate and mjollnir is 72% but that's cus everyone who builds maelstrom will try to turn it into mjollnir by late game and if they are winning they likely will complete it. It does not mean maelstrom is bad, on the contrary it might be MORE efficent gold wise than mjollnir cus mjollnir costs more than twice and the proc isn't twice as good unless it hits all 8 targets.

Maelstrom 0.73
Mjollnir 0.67
Mark if that was true then why doesn't Mjollnir rank higher than Maelstrom?
If it really was 'snowballing' you would expect it to be higher.
What you failed to account for is the other half of the games where you are losing and the high efficiency item brings a comeback.

Most retarded formula ever. You must compare item of same value.

The whole point of this though is to do it for items of different cost.
For the same cost items you could just compare winrates.

Nice theorycrafting,
I feel like pro player now

Nope it's still dumb cus item win rates in dotabuff are very skewed, don't mean much.